azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | | Logs:
juli969 has quit [Quit: Nettalk6 -]
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 265 seconds]
Degi_ is now known as Degi
sorear has quit [Ping timeout: 245 seconds]
sorear has joined #scopehal
<azonenberg> So i just realized
<azonenberg> I'm pretty sure the picoscope 6000 series i have (500 MHz BW, 1.25 Gsps) actually has the performance and channel count to do full protocol decoding and maybe even basic SI measurements on DDR1 RAM
<azonenberg> Might be a nice project
<azonenberg> i want to improve the ram protocol decoding and its an 8 channel scope which is handy
<azonenberg> the little wifi router i used on the stream would be a perfect test subject
<azonenberg> and i could probably port most of the analysis logic to ddr2/3
<azonenberg> things like gating DQ captures by read or write to plot separate read/write eyes
<azonenberg> Also, just swapped some of my wavelink probes around
<azonenberg> and confirmed that the WL-PBUS2 does not work with my SDA
<azonenberg> the same amplifier module works on a WL-PBUS or a WL-PLINK
<azonenberg> When i upgrade the scope this wont be an issue of course
<azonenberg> but short term i might want to pick up another of the old WL-PBUS to be able to use all my probes with it
<azonenberg> Although it seems there's only two on ebay, one of which looks to be in bad shape, and both are ridiculously expensive
sam210723 has quit [Read error: Connection reset by peer]
sam210723 has joined #scopehal
bvernoux has joined #scopehal
<bvernoux> Hello
<bvernoux> Electronic price are totally crazy
<bvernoux> I have just checked STM32F405 which is pretty old and it is now >2 times more expensive than 2months ago !!
<bvernoux> for TCXO and other stuff there is a lead time up to 52weeks ...
<bvernoux> 2021 will be very not fun to build anything ...
sam210723_ has joined #scopehal
sam210723 has quit [Ping timeout: 260 seconds]
sam210723__ has joined #scopehal
sam210723_ has quit [Ping timeout: 246 seconds]
juli969 has joined #scopehal
<Degi> Uhh
<Degi> I measured a pulse with 120 ps risetime on a 350 MHz scope
<Degi> With 1 V amplitude... I wonder what its actual amplitude is
<Degi> Hm I think its more like 350 ps, the pulse itself can be approximated with a sine half wave of 1 ns (for the half wave)
<Degi> Freq measurement says 574 MHz which roughly agrees with my eyes
<azonenberg> bvernoux: yeah MAXWELL is probably going to be on hold for the rest of the year
<azonenberg> because the xc7k160t is out of stock with like a 1-year lead time everywhere
<bvernoux> azonenberg, yes electronic shortage is a real mess
<bvernoux> price explodes and there is no availability ...
<bvernoux> That impact everything not only specialized FPGA, MCU but also TCXO, Synthesizers ...
<bvernoux> It will be great the industry focus to keep some samples to use that time to do advanced prototypes to wait until availability of chipset ;)
<bvernoux> But it is not the case ...
<bvernoux> I'm lucky that I have produced 500+ HydraBus with STM32F405RGT6
<bvernoux> it was just before the rush to buy everything and price is 2x more now ...
<azonenberg> oh i'm sure there are a few samples available if you're a big customer
<bvernoux> yes ;)
<azonenberg> the sales guys are probably just very choosy with who gets them
<bvernoux> yes it's totally crazy
<bvernoux> It is crazy even ST25R3916-AQWT is out of stock worldwide
<bvernoux> as it is ultra specific chipset very new ...
<bvernoux> I was the first in world to use it ;)
<bvernoux> also TCXO better than 1ppm are out of stock worldwide because the factory has burn ...
<bvernoux> it seems like there was only one factory to do all TCXO in world (with different brand which seems to be all built in the same factory)
<azonenberg> So i'm now in talks with Siglent about how they can help out with driver/development support :)
<d1b2> <TiltMeSenpai> whoo 🎉
<d1b2> <mubes> Fix your bloody drivers :-)
<azonenberg> I now have the email and phone contact for a siglent guy who's very interested, we're still chatting back and forth
<d1b2> <mubes> Their scopes have some really nice features esp with respect to things like memory depth....a few tweaks here and there and they would be killers.
<d1b2> <mubes> Very happy to help out if it's useful.
<bvernoux> azonenberg, ha could be a good point to push pressure on Rigol ;)
<bvernoux> azonenberg, I'm in contact with Boris from Rigol EU
<bvernoux> azonenberg, He told me they are working to improve SCPI speed ...
<bvernoux> azonenberg, He work for RIGOL Technologies EU GmbH
<bvernoux> azonenberg, as support from Rigol HeadQuarter does not answer to any Email ...
<bvernoux> azonenberg, last time he has confirmed that with my tool on github and all details they have reproduced the slowness to retrieve data with SCPI commands and they can fully reproduce/analyze it I have even sent them full capture with pcapng to show the issue is on Rigol ...
<d1b2> <mubes> On the siglents the nice thing is that only one driver is needed to cover the 2000/5000/6000 as they are fundementally the same software....there are obviously differences but nothing that can't easily be dealt with via a few switches here and there. With 'factory support' I'd happily help out fixing up the issues which would leave you primarily free to deal with the core...that could actually be a nice model if you could hand off driver support
<d1b2> for each set of scopes to either a factory or community group while you get to concentrate on making 'the whole' better.
_whitelogger has joined #scopehal
<azonenberg> mubes: Yeah so i just got off the phone from a nice productive call with Jason from Siglent, he's very interested in the project and is planning to send me a demo of one of their midrange scopes, needs to figure out exactly what model
<azonenberg> Says sending me a 1 GHz flagship to keep is probably a hard sell but he hopes to be able to send a midrange to keep and arrange a loaner of a flagship for a week or two to test on once we think everything is good
<azonenberg> he's also interested in potentially helping with driver support etc, and just generally loves the idea of the project
<azonenberg> So we'll see how this goes
<d1b2> <TiltMeSenpai> oh wait they're just giving you a midrange Siglent to keep?
<azonenberg> Dont know yet
<azonenberg> he wants to, doesnt have approval to actually do it as of now
<d1b2> <TiltMeSenpai> that's pretty awesome that they're even considering that though
<azonenberg> If someone is willing to commit to being a driver maintainer i could probably arrange to move said scope around if needed
<azonenberg> its not like i need another scope for my own use :p
<azonenberg> But we'll worry about that when the time comes
<azonenberg> So far this looks like it could be a productive relationship though
<azonenberg> i also chatted with him a bit about the possibility of adding the pub/sub API model i use on pico
<azonenberg> where it pushes waveforms out into a fifo and to the socket as fast as it can
<bvernoux> Are you sure Siglent are not as slow as Rigol like MSO5k ?
<bvernoux> as there is really little interest to support a scope with max 1 WFM/s
<bvernoux> (even with 10KSps)
<azonenberg> Not sure, we'll find out
bvernoux has quit [Quit: Leaving]
<Degi> Any idea why the xc7k160t is out of stock?
<azonenberg> Degi: um, the same reason every other chip is hard to get :p
<azonenberg> the global semiconductor shortage
juli969 has quit [Quit: Nettalk6 -]